The Transforming Requirement Sentences in The Form of Requirement Boilerplates to Ontology Using Natural Language Processing

Requirements sentences in the software development process should be consistent. If there is inconsistency, there must be a conflict between requirements in a software project. We propose converting a requirement sentence into an ontology to check such inconsistency. Our approach separates each noun and verb in a requirements sentence using Natural Language Processing (NLP) technique to find concepts (nouns) and relations (verbs) between words. Our proposed ontology structure can be dynamically updated based on changes in the requirements sentences. We validated such an approach using precision and recall testing based on four requirements sentences with three times changes on four requirements sentences. The accuracy results are 100%. It means our approach can accommodate any changes in the requirements.